Particle Filter Based Estimation of Mn-Crust Layers in Sub-Bottom Sonar Data

Cobalt-rich manganese crust (Mn-crust) is a deep sea mineral resource being targeted as a source of critical minerals. Estimating the volumetric distribution of Mn-crust requires measuring its thickness and lateral distribution. Acoustic in-situ methods are proven to be effective for continuous thickness measurements suitable for volumetric estimation. In order to address the challenges in thickness measurement in the presence of noise, this paper proposes a new Bayesian estimator using particle filter in order to detect the reflections in sub-bottom acoustic data. The algorithm developed is applied to field collected data from deep sea surveys of Autonomous Underwater Vehicle BossA. The feasibility of the proposed method is described along with challenges and future research directions.